Github: here
In the United Kingdom, the police have the authority to stop and search individuals based on “reasonable grounds”. The fairness of the process is not fully guaranteed. The project will identify possible biases in the stop-and-search process by analyzing relevant data in 2022 using graphs and maps.
Data used in this project were taken from the official UK police website, the Office for National Statistics, and Wikipedia. The official websites and government agencies have a high degree of credibility and accuracy. Wikipedia is an open editing platform that is not as accurate as the former two, but it provides a wide range of data resources. I acquired the data through the API, direct file downloads, and web scraping, respectively. In addition, I store the data in a SQL database and use SQL for dataset extraction and manipulation. The main calculation used in this project is the stop and search rate, which is calculated as the number of people from a particular group stopped and searched per 1,000 people from the same group.
The following chart analyzes the number of stops per 1,000 people of different races per month, and different colors represent each race.
It is clear that there are significant differences in stop and search rates between ethnicities. Specifically, blacks have the highest stop and search rates in 2022, followed by Asians and other races. Whites and Mixed races have relatively low stop and search rates. It is worth noting that despite the sharp drop in stop rates in November 2022, blacks are still more likely to be stopped relative to other races. This provides the evidence of the existence of racial bias.
The pyramid-shaped bar chart shows the comparison between males and females in the different ethnic groups involved in stop-and-search.
From the graph, when race is not considered, the odds of being stop-and-search are higher per 1,000 males than per 1,000 females. Thus, we can conclude that there is a gender bias in police stopping people. When race is taken into account, we find that both males and females are more likely to be affected by stop-and-search in the black population. Black females, while more likely to be stopped compared to females of other races, the difference between them is not significant. Most significantly, the bar representing black males indicates a significantly higher likelihood of being stopped compared to males of other races.
The pyramid-shaped bar graph below illustrates racial disparities in the likelihood of stop-and-search across different age groups and gender.
Excluding the under-10 age group, it is noteworthy that, across all other age groups, the rate of stop-and-search per 1,000 is much higher for blacks than for other races, regardless of gender. Furthermore, individuals in the 25-34 and 10-17 age groups exhibit the highest likelihood of being stopped and searched, with black males being the most likely to be stopped and searched.
Considering whether this bias varies geographically, the following map shows the odds of being stopped per 1,000 people living in different police force areas, with bars that take race into account attached to each police force area.